Diagnostic accuracy of combining C-Reactive protein and Alvarado Score among 2-to-20-year-old patients with acute appendicitis suspected presenting to Emergency Departments.

OBJECTIVE: Main objective was whether the combination of C-Reactive Protein (CRP) and Alvarado Score (AS) increase the diagnosis accuracy of AS among 2-to-20-year-old patients with suspected acute appendicitis presenting to Emergency Departments.
METHODS: This is a secondary analysis of prospective cohort study consecutively including all patients from 2 to 20 years of age attended for suspected acute appendicitis in 4 Spanish Emergency Departments during 6-month period. We collected demographic, clinical, analytic and radiographic, and surgical data. AS categories were retrospectively calculated as low (0-4 points), intermediate (5-6 points) or high (7-10 points). The cut-off levels were >0.5 mg/dl for CRP. The outcome was diagnosis of acute appendicitis within 14 days of the index visit.
RESULTS: A total of 331 patients with suspected of acute appendicitis (mean age 11.8 (SD 3.8) years; 52.9% males) were recruited. According to AS, 108 (32.6%) were at low risk, 76 at (23.0%) intermediate risk and 147 (44.4%) at high risk of acute appendicitis. One hundred and sixteen (35.0%) cases had confirmed histopathological diagnosis of acute appendicitis. The AUCs of ROC were 0.76 (0.70-0.81) for AS and 0.79 (95% CI 0.75-0.84) for CRP-AS being the difference statistically significant (p=0.003). The CRP for diagnosis acute appendicitis in low risk AS group had negative predictive value of 95.8% (95%CI 87.3-98.9) and likelihood ratio negative of 0.4 (95%CI 0.2-1.0)..
CONCLUSIONS: CRP-AS has shown to increase the diagnostic accuracy of AS for acute appendicitis. This approach may be useful to rule out the diagnosis of acute appendicitis in paediatric patients attended for abdominal pain suggestive of acute appendicitis. ©The Author 2021. Published by Sociedad Española de Quimioterapia. This article is distributed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0)(https://creativecommons.org/licenses/by-nc/4.0/).

INTRODUCCION

Abdominal pain is one of the main reasons for presenting to Emergency Departments (EDs) [1]. Acute appendicitis (AA) is the most common surgical emergency in childhood and one of the main differential diagnoses to consider in paediatric patients and adolescents [2,3].

Despite of advances in the diagnosis and surgical treatment of AA, the diagnosis still remains difficult, especially in paediatric populations [4], being a clinical challenge even for experienced emergency physicians [5]. The clinical history and physical examination may be less useful compared to adults because children are more frequently associated with atypical clinical manifestations [6]. Differential diagnosis of an AA is extensive, and many clinical conditions can mimic AA [5]. The delay in the diagnosis increases morbidity and mortality, whereas false positive diagnosis of AA leads to unnecessary surgery. In this sense, the studies show a frequency of perforation between 17 and 33% and a percentage of negative laparotomy between 3 and 54% [2].

Different clinical scores have been developed to help physicians to improve the accuracy of the diagnosis of AA, reducing the time needed for diagnosis and the number of inappropriate imaging tests and appendectomies [7]. The Alvarado score (AS) is the most frequently score used in clinical practice [8]. The AS considers symptoms, signs, and laboratory, which are leukocytosis and neutrophilia (Table 1). Some studies have shown that diagnostic accuracy of AS is not sufficient to establish the presume diagnosis or rule out, especially in children, and this can lead to delays in the diagnosis increasing the risk of appendicitis perforations [7,8].

Table 1

Alvarado Score for Appendicitis Diagnosis

Manifestations	Value
Signs
Right Lower Quadrant Tenderness	2
Elevated Temperature (37.3°C or 99.1°F)	1
Rebound Tenderness	1
Symptoms
Migration of Pain to the Right Lower Quadrant	1
Anorexia	1
Nausea or Vomiting	1
Laboratory Values
Leukocytosis > 10,000 mm3	2
Leukocyte Left Shift	1

Risk of acute appendicitis: low (0-4 points), intermediate (5-6 points) or high (7-10 points)

The C-reactive protein (CRP) is a protein synthetized from hepatocytes that increases within 4-6 hours after acute tissue injury and peaking at 36-48 hours. One hand, it has been reported that CRP used solely or in combination with white blood cell (WBC) could be effective in the discrimination between acute and complicated appendicitis [9]. On the other hand, there is little evidence and the results are contradictory among paediatric patients and adolescents [10,11]. Despite this, some authors recommend evaluating routinely CRP level, WBC and Neutrophil percentage (NP) among patients with initial diagnosis of appendicitis [12,13].

Considering all previously written, further evidence is necessary to confirm the role of CRP in conjunction with AA in the diagnosis of acute appendicitis among paediatric patients and adolescents. In this sense, the main objective was to determinate whether the combination of CRP and AS increase the diagnosis accuracy of AS among 2-to-20-year-old patients with acute appendicitis suspected presenting to EDs. The secondary objectives were to study the diagnostic accuracy of the CRP level in the different categories of AS in this group of patients and to compare the diagnostic capacity of CRP level with WBC and NP.

MATERIAL AND METHODS

Study design and setting. We performed a secondary analysis of observational prospective cohort study that consecutively included children and adolescents attended for abdominal pain suspected of AA in 4 Spanish ED (Hospital Clínico San Carlos of Madrid, Hospital de Basurto of Bilbao, Hospital Virgen de la Macarena of Sevilla and Hospital Sant Joan de Déu of Barcelona) from June to December 2014. The study was approved by the Ethical Committee of the reference centre and was carried out according to the principles of Good Clinical Practice. Informed consent was obtained from the adolescents and from the parents of the children included in the study.

Population. All patients from 2 to 20 years of age with abdominal pain suspected of AA of less than 72 hours of evolution were consecutively included in the study. The exclusion criteria were: patients with a radiological test of the abdomen prior to attending the ED, pregnancy, history of appendicectomy, inflammatory disease or active cancer, abdominal traumatism, surgery or invasive abdominal procedure within the previous 7 days, use of systemic steroids in the last 14 days, receipt of any other immunosuppressive treatment or chemo-therapy within the previous 29 days and participation in an investigation study in the previous 30 days.

Study protocol. The patients were assessed by the attending physician following the standard protocol of usual clinical practice, that is, collecting the clinical history and performing a physical examination and routine laboratory tests and, if necessary, an imaging test, ultrasonography and/ or computerized tomography, and evaluation by a surgeon on duty. The laboratory and imaging tests and interconsultation with an on-duty surgeon were requested as per the criteria of the attending physician, independently of the participation in the study. In patients undergoing appendicectomy, the diagnosis of AA was made from the surgical piece by a specialist in anatomopathology. All the patients discharged directly without surgery were evaluated at 14 days to know whether they had received medical care and had a clinical and/or histo-pathological report compatible with AA.

Variables. The investigators of each centre collected the following data on a standardized form: demographic (age and gender), clinical (symptoms, signs and time of evolution), analytical (WBC, NP and CRP) and radiological (abdominal ultra-sonography and/or computerized tomography) and, if necessary, surgical and histopathological data. The forms were reviewed by the coordinating investigator of each centre. The patients were retrospectively classified according to the AS as low (0-4 points), intermediate (5-6 points) or high (7-10 points) risk of AA. The cut-off levels were > 0,5 mg/dl for CRP, ≥10 x 109 /L for WBC and ≥ 75% for NP. The final diagnosis of AA was based on the report of the histopathological study of the resected appendix in patients who had undergone appendicectomy, or via a telephone call at 2 weeks of follow up to know if they had been histologically diagnosed. AA was histologically demonstrated by mucosal neutrophil infiltration of the appendix with or without local peritonitis.

Statistical analysis. Categorical variables are expressed as numbers and percentages and the quantitative variables are expressed as means and standard deviations or medians and interquartile ranges if the distribution was not normal (this was tested using Kolmogorov-Smirnov test). Categorical variables were compared with the Pearson chi-square test or Fisher test and quantitative variables using the Student’s-t test (or the Mann-Whitney U test if the distribution was not normal). Logistic regression was used to combine CRP and AS in the diagnosis of acute appendicitis. The values of sensitivity (Se), specificity (Sp), positive predictive value (PPV) and negative predictive value (NPV), positive likelihood ratio (LHR+) and negative likelihood ratio (LHR-) were calculated for CRP and CRP stratified by AS. To determine the discriminatory capacities of the CRP, leukocytosis, neutrophilia, AS and combination of AS and CRP to diagnose acute appendicitis the areas under the curve (AUC) of receiver-operating characteristic (ROC) curves were calculated. The AUC of ROC were compared by DeLong test. We considered differences to be statistically significant if the p value was less than 0.05 (two-tail), and 95% CI of AUC ROC excluded 0.5. Statistical analyses were performed using SPSS 24.0 (SPSS Inc., Chicago, IL) STATA 14.0 statistical package (StataCorp. 2015. Stata Statistical Software: Release 14. College Station, TX: StataCorp LP).

RESULTS

Patient characteristics. We included 331 patients with suspected of AA. The mean age was 11.8 (SD 3.8) years and 175 (52.9%) were males. Table 2 shows the characteristics of the patients included in the study. The median time of clinical course was 24 (IQR 9–48) hours. According to AS, 108 (32.6%) were at low risk, 76 at (23.0%) intermediate risk and 147 (44.4%) at high risk of AA. At least one imaging test was performed in 189 (57.1%): 172 (52.0%) ultrasonographies, 6 (1.8%) CTs and both in 11 (3.3%) cases. Appendectomy was carried out during the index visit in 118 (35.6%) patients, 109 (94.0) of whom had a histopathological diagnosis of AA. At the end of the 2-week follow-up, 116 (35.0%) had confirmed histopathological diagnosis of AA.

Table 2

Characteristics of the patients included in the study based on the diagnosis of appendicitis.

	All Patients(n=331)	Appendicitis(n=116)	No Appendicitis(n=215)	p
Demographic data
Age (years); [mean (SD)]	11.8 (3.8)	11.1 (3.7)	12.2 (3.8)	0.013
Male; [n (%)]	175 (52.9)	74 (63.8)	101 (47.0)	0.003
Clinical data
Time since symptom onset (hours); [n (%)]				<0.001
0-12	88 (26.6)	19 (16.4)	69 (32.1)
12-24	63 (19.0)	30 (25.9)	33 (15.3)
24-48	92 (27.8)	42 (36.2)	50 (23.3)
48 o more	88 (26.6)	25 (21.6)	63 (29.3)
Associated Symptoms; [n (%)]
Periumbilical pain with migration to LRQ	286 (86.4)	104 (89.7)	182 (84.7)	0.205
Anorexia	110 (33.2)	53 (45.7)	57 (26.5)	<0.001
Nausea	165 (49.8)	69 (59.5)	96 (44.7)	0.100
Vomiting	139 (42.0)	65 (56.0)	74 (34.4)	<0.001
Physical Examination; [n (%)]
Fever (37.8ºC)	35 (10.6)	15 (12.9)	20 (9.3)	0.306
LRQ Tenderness	290 (87.6)	109 (94.0)	181 (84.2)	0.010
Rebound tenderness	190 (57.4)	84 (72.9)	106 (49.3)	<0.001
Percussion tenderness	200 (60.4)	87 (75.0)	113 (52.6)	<0.001
Rigidity	98 (29.6)	52 (44.8)	46 (21.4)	<0.001
Rovsing sign	53 (16.0)	26 (22.4)	27 (12.6)	0.020
Laboratory data
White blood cell count (ml/mm3); [mean (SD)]	11.7 (6.0)	14.5 (4.7)	10.2 (6.2)	<0.001
Neutrophile (mil/mm3)	8.5 (4.9)	11.4 (4.7)	6.9 (4.2)	<0.001
C-reactive protein (mg/L); [median (IQR)]	0.83 (0.29-4.08)	3.02 (0.61-6.57)	0.30 (0.19-1.96)	<0.001
Alvarado Score; [n (%)]
Low-risk	108 (32.6)	11 (9.5)	97 (45.1)	<0.001
Intermediate-risk	76 (23.0)	18 (15.5)	58 (27.0)
High-risk	147 (44.4)	87 (75.0)	60 (25.9)
Imaging data
US; [n (%)]	172 (52.0)	58 (50.0)	114 (53.0)	0.599
CT; [n (%)]	6 (1.8)	3 (2.6)	3 (1.4)	0.438
US and CT; [n (%)]	11 (3.3)	5 (4.3)	6 (2.8)	0.462
No imaging; [n (%)]	189 (57.1)	66 (56.9)	123 (57.2)	0.956
Treatment data
Surgery; [n (%)]	118 (35.6)	109 (94.0)	9 (4.2)	<0.001

SD: standard deviation; LRQ: lower right quadrant; IQR: interquartile range; US: ultrasound; CT: computerized tomography

Table 2 shows the characteristics of the sample and the univariate analysis based on the presence or not of appendicitis. Statistically significant differences were observed in the age, gender, time of evolution, associated symptoms and physical examination, WBC, NP, CRP levels, the risk categories according to the AS, and the performance of exploratory surgery.

Primary data analysis. The figure 1 shows the diagnosis capacity of combination CRP and AS, and AS. The AUCs of ROC were 0.76 (0.70-0.81) for AS and 0.79 (95% CI 0.75-0.84) for CRP-AS being the difference statistically significant (p=0.003).

Figure 1

Comparations between Area under curve (AUC) ROC of leukocytosis, neutrophilia, C-reactive protein (CRP), Alvarado Score (AS) and the combination of CRP and AS for the diagnosis of acute appendicitis.

Secondary data analysis. The table 3 shows the characteristic of CRP test along and stratified between three AS groups. Six (5.2%) out of 116 patients with AA were classified as low risk AS and CRP ≤ 0.5 mg/dl. The CRP for diagnosis AA in low risk AS group had NPV of 95.8% (95% CI 87.3-98.9) and LHR- of 0.4 (95% CI 0.2-1.0). A negative CRP result reduced the probability of AA from 8.8% to 5.2% in this low risk AS group. Regarding the comparation between diagnostic accuracy of CRP level with WBC and NP, no statistically significant differences were found.

Table 3

Diagnosis characteristic of Alvarado Score and C-Reactive Protein test and C-Reactive Protein stratified between three Alvarado Score groups.

	Se(95% CI)	Sp(95% CI)	PPV(95% CI)	NPV(95% CI)	LHR+(95% CI)	LHR-(95% CI)	AUC(95% CI)
Alvarado score threshold ≥5(low vs intermediate-high)	90.5 (83.3-94.9)	45.1 (38.4-52.0)	47.1 (40.4-53.8)	89.8 (82.1-94.6)	1.7 (1.4-1.9)	0.2 (0.1-0.4)	0.68 (0.62-0.74)
Alvarado score threshold ≥7(low-intermediate vs high)	75.0 (65.9-82.4)	72.1 (65.5-77.9)	59.2 (50.8-67.1)	84.2 (78.0-89.0)	2.7 (2.1-3.4)	0.35 (0.25-0.48)	0.73 (0.68-0.79)
CRP (Cut-off 0.5 mg/dl)	79.3%(70.6-86.1)	56.7(49.8-63.4)	49.7(42.3-57.1)	83.6(76.3-89.0)	1.8(1.5-2.2)	0.4(0.3-0.5)	0.68(0.62-0.74)
CRP stratified by Alvarado score
Low risk
CRP (Cut-off 0.5 mg/dl)	72.7%(39.3-92.7)	70.1%(59.8-78.8)	21.6%(10.4-38.7)	95.8%(87.3-98.9)	2.4(1.5-3.9)	0.4(0.2-1.0)	0.71(0.55-0.87)
Intermediate risk
CRP (Cut-off 0.5 mg/dl)	72.2%(46.5-89.3)	50.0%(36.7-63.3)	31.0%(18.1-47.2)	85.3%(68.2-94.5)	1.4(1.0-2.1)	0.6(0.3-1.2)	0.61(0.46-0.76)
High risk
CRP (Cut-off 0.5 mg/dl)	81.6%(71.6-88.8)	41.7%(29.3-55.1)	67.0%(57.1-75.6)	61.0%(44.5-75.4)	1.4(1.1-1.8)	0.4(0.3-0.8)	0.62(0.52-0.71)

Se: Sensibility; Sp: Specificity; PPV: Positive predictive value; NPV: Negative predictive value; LHR+: Likelihood ratio positive; LHR-: Likelihood ratio negative; AUC: area under curve (ROC). CRP: C-reactive protein.

DISCUSSION

The present study shows that the combination of AS and CRP is a useful diagnostic tool to rule out the diagnosis of AA in patients from 2 to 20 years of age attended for abdominal pain suggestive of AA. This strategy showed an improvement in the diagnostic accuracy of the AS used alone.

These findings increase the evidence about the limited diagnosis accuracy of AS to be used in clinical practice as a unique strategy for surgery decision making among paediatrics patients with suspected AA [14-17]. Although the utility of inflammatory markers is still controversial in children, the combination of AS and CRP improved the diagnostic capacity compared with the AS alone [17]. This strategy was especially helpful in low risk AS category.

The CRP, WBC and NP have been the serum markers more studies [18]. Our data confirm that these biomarkers have poor test characteristics alone [17,19]. Regarding the CRP, previous studies have been focused on the diagnosis of AA or differential diagnosis between early and advanced AA [17]. In this sense, a higher cut-off (>3mg/dl) of CRP than our study (>0,5mg/dl) was used [20]. Our strategy was opposed, to increase the negative predictive value of the low-risk AS category, adding the CRP. In fact, the estimated probability of AA varied from 15.7% to 5.2% depending on CPR value in low-risk AS category.

The combination of AS, which includes WBC and NP, allowed identifying a low-risk subgroup of paediatric patients in which more conservative management may be applied with imaging examinations (avoiding or delaying) or final destination (hospital observation or close follow-up if discharged). We agree with other authors that the intent of the model is not to establish the diagnosis of AA, if not to reduce the level of uncertainty [21].

A recent study carried on 402 patients with 8.5+/-3 (range 3 to 13,8) years admitted for suspected AA showed a negative predictive value of 79.7% for AS + CRP and 78.3% for AS + CRP + US, using a cut-off point of 0.6 mg/dL for CRP, among low-risk AS category [22]. Other recent study with 200 patients, observed more controversial results regarding to accuracy of these clinical scores [23]. In our research, a lower cutoff point (<=0,5mg/dl) showed a negative predictive value of 95.8% (87.3%-98.9%) in low-risk AS category despite the time since symptom onset was lower in our cohort. Another study conducted in adults with suspected of AA, showed a negative predictive value of 86%for CRP (cut-off 0.5 mg/dl) in low-risk AS category [24].

Other scoring systems have included the CRP such as Appendicitis Inflammatory Response Score (AIR Score) [25] and INFURG-SEMES score [26]. A recent study that compared AIR Score with AS showed an improvement of the diagnostic capacity and a higher NPV (95% vs 90% in low-risk vs intermediate-high risk groups) [25]. Thus, panels of biomarkers, such as APPY 1 Test that also includes CRP, are being researched to rule out the AA in EDs [27].

The present study presents several limitations. First, those inherent in the design of the study. Second, the time of course of AA could have modified the findings of the study. In this sense, the sample size was limited to do stratified analyses based on the time of sample obtainment for the CRP at arrival to ED concerning the onset of the symptomatology. Third, this was an observational study and, thus, we cannot demonstrate that the addition of CRP to AS reduces the number of imaging tests, unnecessary surgical procedures, the times of hospital stay or the associated costs. The fourth limitation is a possible detection bias in the AA on the index visit. To avoid this, we used a 2-week follow-up for patients discharged without surgery to confirm they were correct classification. Finally, we did not evaluate interobserver agreement in the AS calculation because this is a validated model.

In conclusion, the CRP is a biomarker extensively available in EDs. The combination of CRP and AS has shown to increase the diagnostic accuracy of AS for AA. This approach may be useful to rule out the diagnosis of AA in patients from 2 to 20 years of age attended for abdominal pain suggestive of AA.

Table 4

Estimated probability of acute appendicitis according to the risk categories of Alvarado Score and positive C-reactive protein.

Alvarado score	Probability of AA (%)	CRP (mg/dl)	Probability of AA (%)
Low risk	8.8	≤0.5	5.2
		> 0.5	15.7
Intermediate risk	27.6	≤0.5	15.2
		> 0.5	37.7
High risk	58.2	≤0.5	36.9
		> 0.5	66.4
TOTAL	35.0	≤0.5	16.4
		> 0.5	49.7

AA: acute appendicitis; CRP: C-reactive protein